Deep Learning HSGP4: Hyperparameters analysis

  1. Edna Segura 1
  2. Hans Carrillo 1
  3. Rosario López Gómez 1
  4. Iván Pérez 1
  5. Montserrat San-Martín 2
  6. Juan F. San Juan 1
  1. 1 Universidad de La Rioja
    info

    Universidad de La Rioja

    Logroño, España

    ROR https://ror.org/0553yr311

  2. 2 Universidad de Granada
    info

    Universidad de Granada

    Granada, España

    ROR https://ror.org/04njjy449

Actas:
SPACEFLIGHT MECHANICS 2021: Proceedings of the 31th AAS/AIAA Space Flight Mechanics Meeting held February 1–3, 2021, Virtual Event
  1. Carolin Frueh (ed. lit.)
  2. Renato Zanetti (ed. lit.)
  3. Jeffrey R. Stuart (ed. lit.)
  4. Angela L. Bowes (ed. lit.)

Editorial: Univelt Inc.

ISSN: 0065-3438

ISBN: 978-0-87703-679-1

Año de publicación: 2021

Volumen: 176

Páginas: 1299 - 1313

Congreso: 31th AAS/AIAA Space Flight Mechanics Meeting. February 1–3, 2021, Virtual Event

Tipo: Aportación congreso

beta Ver similares en nube de resultados
GOOGLE SCHOLAR
Repositorio institucional: lockAcceso abierto Editor

Resumen

The hybrid orbit propagation methodology is used to model the error of any type of orbitpropagator with the aim of improving its perturbation model or integration technique andhence enhancing its accuracy. In this work, we present an application of the hybrid meth-odology, in which the time-series forecasting process is performed using deep learningmethod to SGP4. We have adjusted the resulting Hybrid SGP4 propagator, HSGP4, to thecase of Galileo-type orbits. We will describe the hyper-parameter selection, which is animportant part of the development of HSGP4, and show how HSGP4 can improve theaccuracy of SGP4.